Facial trauma and the risk of intracranial injury in motorcycle riders

Facial trauma and the risk of intracranial injury in motorcycle riders

INJURY PREVENTION AND TRAUMA/ORIGINAL RESEARCH Facial Trauma and the Risk of Intracranial Injury in Motorcycle Riders Jess F. Kraus, MPH, PhD Thomas...

104KB Sizes 0 Downloads 28 Views

INJURY PREVENTION AND TRAUMA/ORIGINAL RESEARCH

Facial Trauma and the Risk of Intracranial Injury in Motorcycle Riders

Jess F. Kraus, MPH, PhD Thomas M. Rice, MPH Corinne Peek-Asa, MPH, PhD David L. McArthur, PhD, MPH From The Southern California Injury Prevention Research Center, UCLA School of Public Health, Los Angeles, CA. Dr. Peek-Asa is now with the Injury Prevention Research Center, University of Iowa College of Public Health, Iowa City, IA. Dr. McArthur is now with the Division of Neurosurgery, UCLA School of Medicine, Los Angeles, CA.

Study objective: We describe the associations among facial fracture, helmet use, skull fracture, and traumatic brain injury in injured motorcycle riders. Methods: The study population consisted of 5,790 motorcycle riders who sustained an injury from a crash in 1991, 1992, or 1993 and were identified from emergency department or hospital records in 28 hospitals in 11 California counties. Diagnostic information was abstracted from ED or hospital records and coded to the 1990 Abbreviated Injury Scale. The risk of traumatic brain injury was examined by using odds ratios and 95% confidence intervals. Logistic regression analysis was used to examine the associations among helmet use, skull fracture, facial fracture, and traumatic brain injury. Results: Facial injury was diagnosed in 24.4% of all injured riders, including 411 with one or more facial fractures. The odds of traumatic brain injury were 3.5 times greater with than without a facial injury and 6.5 times greater with a facial fracture than with no facial injury. Significantly increased odds of traumatic brain injury were observed for fracture of all bones of the face, but the highest odds of traumatic brain injury were found in riders with fractures to bones of the upper face. Helmet use status and the presence of skull fracture were found to be significant effect modifiers. Facial fracture with concurrent skull fracture increased the risk of traumatic brain injury dramatically. Facial fractures are more strongly associated with traumatic brain injury in helmeted riders. Conclusion: The presence of facial fractures increases the risk of traumatic brain injury. Riders with facial fractures should be screened for brain injury, regardless of helmet use status. [Ann Emerg Med. 2003;41:18-26.]

Copyright © 2003 by the American College of Emergency Physicians. 0196-0644/2003/$30.00 + 0 doi:10.1067/mem.2003.1

1 8

ANNALS OF EMERGENCY MEDICINE

41:1

JANUARY 2003

TRAUMA IN MOTORCYCLE RIDERS Kraus et al

INTRODUCTION

Facial injury, particularly soft tissue injury and fracture of the facial bones, is a frequent result after motor vehicle crashes, falls, and assaults1-5 and crashes during some recreational activities, such as bicycling6 and skiing.7 Data from the Major Trauma Outcomes Study,2 with information on more than 87,000 patients in the United States in 1995, show that more than 34% of the patients in the study had facial injuries, and 25% of these had a facial fracture. Although the literature on the medical and surgical management of facial trauma is extensive,8-10 the association of facial injury with other injuries, especially intracranial trauma, is infrequently reported. Lee et al11 report that the risk of closed head injury among patients with facial injury was greatest with fractures to the bones of the upper face, whereas patients with only fractures of the mandible or fractures to the midface region were less likely to have concurrent head injuries. Davidoff et al12 found that 55% of patients with facial fractures had a closed head injury. However, Haug et al13 reported only a 17.5% incidence of closed head injuries in patients with facial fractures. Nakhgevany et al1 report that 68% of those with facial trauma had associated head injuries, but the nature of the facial injury and details of the head injuries were not provided. More recently, Keenan et al14 evaluated the risk of traumatic brain injury with facial fractures and concluded that, although the overall risk of traumatic brain injury was increased for all patients with facial bone fractures, the highest risk of traumatic brain injury was found for those with fractures of the maxilla. Unfortunately, these either did not provide estimates of risk of traumatic brain injury with facial fracture or had sample sizes that were too small for risk estimates (odds ratios [ORs]) for several of the facial bones. The vulnerability of motorcyclists and bicyclists to facial injury in crashes has prompted several articles in which the role of helmets (and their design) in reducing facial injury risk was examined. All reports3,4,6,15-17 confirm the protective role of helmets in reducing the occurrence of facial injuries in motorcycle and bicycle

JANUARY 2003

41:1

ANNALS OF EMERGENCY MEDICINE

riders and that the full-face model, as compared with the open-face model, was superior in terms of injury attenuation.3,15 However, none of the published reports examined traumatic brain injury, facial injuries, and helmet use status simultaneously. Most reports on the relationship of facial injury to intracranial trauma11-14,18-21 suffer from shortcomings that dampen the utility of their findings. The most common problems are failure to use an adequate comparison group, small patient samples, lack of specificity of exposure (ie, number, severity, and location of the fractured facial bones), lack of specificity of outcome (ie, nature and severity of intracranial injury), absence of measures of risk, and failure to adjust for potential confounders when estimating the association between facial injury and traumatic brain injury. This report addresses many of these problems and examines data from 5,790 patients injured in motorcycle crashes to compare the risks of intracranial injury with and without specific facial bone fractures, considering both helmet use and skull fracture. M AT E R I A L S A N D M E T H O D S

Selected details of the study design, case finding, data abstraction, injury coding, and analyses specific to this report are included here. Full details of the study parameters are found in an earlier publication.4 Institutional review board approval was granted for the original study protocol dated 1991. The research design was a cohort study consisting of injured motorcycle riders treated in 28 hospitals located in the California counties of Contra Costa, Fresno, Kern, Los Angeles, Orange, Riverside, Sacramento, San Bernardino, San Diego, Santa Clara, and Shasta. These counties accounted for 65% of the motorcycle crash fatalities and 71% of the state’s population in 1990. Patients were included for study if their injuries occurred between January 1, 1991, and December 31, 1993. Hospitals were selected on the basis of geographic location, range of treatment level, funding type, bed number, and service to different patient profiles. Of these hospitals, 22 were Level I or II trauma centers; the

1 9

TRAUMA IN MOTORCYCLE RIDERS Kraus et al

remaining 6 had emergency departments but were not trauma centers. Hospital records were reviewed if any of the following sources indicated motorcycle crash involvement: ED logs, radio dispatching logs, emergency medical services (EMS) transportation records, and external cause of injury codes (E-codes) from hospital or ED discharge summaries. All injured riders admitted to any 1 of the 28 hospitals were included, and those treated in the ED and released were included from 9 of the 28 hospitals. All relevant hospital admission and ED records were individually reviewed. Injured riders declared dead at the scene were excluded from the study cohort because they were not admitted to a hospital and might not have had complete diagnostic testing. All injury diagnoses found in ED or hospital records were coded to the 1990 version of the Abbreviated Injury Scale,22 a comprehensive taxonomy of individual injuries. The Abbreviated Injury Scale divides injuries into body region, type of structure, nature, and severity. The Abbreviated Injury Scale uses 9 body regions, including head, chest, abdomen, spine, neck, face, upper extremity, lower extremity, and external. Severity codes range from 0 (no injury) to 6 (unsurvivable injury). The Maximum Abbreviated Injury Severity score denotes the most severe injury in each of the 9 regions. All injury codes were derived by 2 clinically trained staff members. For the purposes of this report, “exposure” was defined as the presence of 1 or more physician-diagnosed facial fractures and included the following Abbreviated Injury Scale codes: mandible (2506XX), maxilla (2508XX), nose (2510XX), orbit (2512XX), and zygoma (2518XX). As has been the practice in most articles dealing with facial fracture and traumatic brain injury,1-3,5,6,11,13,16-21 patients having only fractures, dislocation or avulsion of the teeth (2514XX), or sprain or dislocation of the temporomandibular joint (2516XX) were not included in the exposed group. Fractures to the alveolar ridge (2502XX) were included with fractures to the maxilla. Facial fracture diagnoses without an explicit bone are classified as “facial fracture NFS” (not further specified) after Abbreviated Injury Scale guidelines and were included only in Table 1 and excluded from further analyses.

2 0

“Outcome” was defined as physician-diagnosed traumatic brain injury and included the following Abbreviated Injury Scale injury descriptions and codes: closed head injury NFS (115099), penetrating injury (11600X), traumatic brain injury NFS (115299), injury to any intracranial artery or sinus (12XXXX), brainstem (1402XX), cerebellum (1404XX), cerebrum (1406XX), and pituitary (140799). Unconsciousness in the absence of any of the aforementioned diagnoses (160XXX) and cerebral concussion (161000) were also included as outcomes. Skull fracture (150XXX) in the absence of any of the aforementioned diagnoses was excluded as a traumatic brain injury outcome. After data abstraction, coding, and quality control edits, numeric and percentage distributions were computed. Relative risks of traumatic brain injury were estimated by using exposure ORs and 95% confidence intervals (CIs). ORs were derived for facial region (upper versus lower), specific facial bones, and number of different facial bone fractures. In each of the tabular analyses, the traumatic brain injury ORs were derived for riders wearing and not wearing motorcycle helmets. ORs and CIs were not calculated for comparisons having fewer than 4 patients in any cell. Logistic regression models were used to estimate the main effects of facial fractures on traumatic brain injury occurrence and to examine the interaction effects of the key variables of facial fracture, skull fracture, and helmet use on traumatic brain injury occurrence. Each model was fitted separately for helmeted and unhelmeted riders because simple stratified analyses demonstrated that the facial fracture effect estimates varied by helmet use status. Maximum likelihood estimates were obtained by using the LOGISTIC procedure in the SAS/STAT software package (version 8, SAS Institute, Inc., Cary, NC). Motorcycle riders for whom helmet use status could not be determined were excluded from analyses that involved stratification on this factor. Logistic regression models were specified by using a priori knowledge of potential confounders, effect modifiers, or both. Model selection did not use significancebased procedures or any other selection algorithm. In each regression, the probability of a traumatic brain injury outcome was modeled. Variables in model 1 in-

ANNALS OF EMERGENCY MEDICINE

41:1

JANUARY 2003

TRAUMA IN MOTORCYCLE RIDERS Kraus et al

cluded mandibular fracture, maxillary fracture, nasal fracture, orbital fracture, zygomatic fracture, skull fracture, and all skull fracture–facial fracture product terms. Variables in model 2 included upper facial fracture (nasal, orbital, or zygomatic fracture), lower facial fracture (mandibular or maxillary fracture), skull fracture, upper-lower facial fracture product term, and all skull fracture–facial fracture product terms. Each model was fitted separately to data from helmeted and unhelmeted riders. Model adequacy was assessed with the likelihood ratio test, the Hosmer-Lemeshow goodness-of-fit statistic (χ2HL),23 and the Pearson χ2 statistic. Regression diagnostic procedures included the assessment of each observation’s influence on the Pearson χ2 statistic and the plotting of Cook distances against predicted probabilities. R E S U LT S

Facial injuries, including fractures and soft tissue damage, were diagnosed in 1,410 (24.4%) of the total cohort of 5,790 injured motorcycle riders. Facial fractures

Table 1.

Selected characteristics of injured motorcycle riders, California, 1991 to 1993. TBI Present, No. (%)

Variable

TBI Not Present, No. (%)

Figure.

Percentage distribution of facial fractures in injured motorcycle riders, California, 1991 to 1993.

Sex Male Female Mean age, y Facial injury Yes No Facial fracture Yes No Skull fracture Yes No Helmeted Yes No Total

1,456 (90.0) 162 (10.0) 28.7

3,778 (90.6) 394 (9.4) 29.0

687 (42.5) 931 (57.5)

723 (17.3) 3,449 (82.7)

262 (16.2) 1,356 (83.8)

149 (3.6) 4,023 (96.4)

320 (19.8) 1,298 (80.2)

70 (1.7) 4,102 (98.3)

722 (46.3) 839 (53.7) 1,618

2,152 (65.6) 1,182 (34.4) 4,172

41:1

Orbit 16.8%

Face not further specified 5.6%

Nasal 16.8% Zygoma 15.6% Maxilla 23.3% Mandible 22.0%

TBI, Traumatic brain injury.

JANUARY 2003

were diagnosed in 411 (29.1%) of the 1,410 riders with any facial injury. The mean age of injured riders was 28.7 years when traumatic brain injury was present and 29.0 years when traumatic brain injury was not present. Approximately 90% of injured riders were men. There were no significant differences in the age or sex distributions in riders with and without facial injuries, in riders with and without facial fractures, or in riders with and without soft tissue facial injury. Table 1 summarizes characteristics of the study population on the basis of traumatic brain injury status. The facial bones most frequently fractured were the maxilla (23.3%) and mandible (22.0%; Figure). The zygoma was less frequently fractured (15.6%). The odds of a traumatic brain injury for injured motorcycle riders with a concurrent facial injury of any type were 3.52 times greater (95% CI 3.09 to 4.01) than for those without a facial injury. The chances of a traumatic brain injury increased dramatically for those with a facial bone fracture (OR 6.51; 95% CI 5.23 to 8.11) compared with those with no facial injury. In addition, the odds of a traumatic brain injury were significantly increased (OR 2.74; 95% CI 2.36 to 3.18) for those with only soft tissue injuries to the face. The distribution of the number of fractures to each specific bone is given in Table 2. More than 90% of 624 bones fractured had only 1 fracture within each bone,

ANNALS OF EMERGENCY MEDICINE

2 1

TRAUMA IN MOTORCYCLE RIDERS Kraus et al

and 9.1% had more than 1 fracture per bone. The mandible and maxilla were more likely to have multiple fractures (15.1% and 14.3%, respectively) than the zygoma (1.0%). The association of facial fracture with traumatic brain injury occurrence varied by helmet use status. The OR for traumatic brain injury was 5.13 (95% CI 3.57 to 7.37) in riders wearing helmets and 3.29 (95% CI 2.49 to 4.33) in riders not wearing a helmet. The difference in ORs was marginally significant (Woolf test of homogeneity: χ2=3.68, P=.055); thus, the remaining analyses treated helmet use status as a potential effect modifier. Higher odds of traumatic brain injury are seen for riders with 2 or more fractured facial bones (OR 6.09; 95% CI 4.42 to 8.40) than for those with only one fractured facial bone (OR 4.43; 95% CI 3.39 to 5.78). The odds of traumatic brain injury are related to the number and location of bones fractured (Table 3, right columns). Among riders with only 1 fractured facial bone, the highest odds of traumatic brain injury (OR 5.93; 95% CI 2.64 to 13.61) are seen for those with an orbital bone fracture. Among riders with 2 or more fractured facial bones, the highest ORs for traumatic brain injury are seen for the zygoma or orbit (OR 7.69 and 7.48, respectively).

Table 2.

Number of motorcycle riders with facial fractures by number of fractures to specific bones and percentage with more than 1 fracture per bone, California, 1991 to 1993. No. of Fractures per Bone Fractured Bone Mandible Maxilla Nose Zygoma Orbit All bones *Each

1

2

3

4

Total*

% With >1 Fracture

124 132 109 101 101 567

20 14 2 1 8 45

1 7 — — 2 10

1 1 — — — 2

146 154 111 102 111 624†

15.1 14.3 1.8 1.0 9.9 9.1

row total represents motorcycle riders with ≥1 fractures in each bone. Motorcycle riders might appear more than once. number of fractured facial bones in all motorcycle riders.

†Total

2 2

The likelihood of traumatic brain injury with fractured facial bones varies by helmet use, number of bones fractured, and fracture location (Table 3). Among riders with 1 or more fractured facial bones, the ORs for traumatic brain injury are generally higher for those wearing a helmet. Although 174 motorcycle riders had fractures of 2 or more facial bones, there were too few patients in many of the cells involving 2-bone combinations stratified by helmet use, and therefore, meaningful analyses could not be undertaken. Motorcycle riders with at least one Maximum Abbreviated Injury Severity level 2 or at least one Maximum Abbreviated Injury Severity level 3 facial fracture have higher odds (OR 7.72 [95% CI 5.84 to 10.20] and OR 6.67 [95% CI 3.40 to 13.20], respectively) of traumatic brain injury compared with those having only Abbreviated Injury Scale level 1 fractured bone of the face (OR 5.48; 95% CI 4.14 to 7.25). The fitted models were as follows. In model 1, unhelmeted riders had a likelihood ratio χ2 value of 249.3 (df=10, P<.0001), a Pearson χ2 value of 1,996.4 (df=1,997), and a χ2HL value of 2.4 (df=3, P=.49), and helmeted riders had a likelihood ratio χ2 value of 176.7 (df=10, P<.0001), a Pearson χ2 value of 2,859.9 (df=2,860), and a χ2HL value of 3.3 (df=3, P=.35). In model 2, unhelmeted riders had a likelihood ratio χ2 value of 256.3 (df=6, P<.0001), a Pearson χ2 value of 1,996.6 (df=1,997), and a χ2HL value of 0.01 (df=3, P=1.00), and helmeted riders had a likelihood ratio χ2 value of 180.4 (df=6, P<.0001), a Pearson χ2 value of 2,871.0 (df=2,860), and a χ2HL value of 0.36 (df=3, P=.95). Regression diagnostic procedures confirmed the adequacy of the models and did not identify any overly influential observations. Table 4 shows crude and adjusted ORs from logistic regression model 1. Each adjusted measure estimates the effect of the facial fracture on traumatic brain injury, independent of all other facial bone and skull fractures. The factor with the highest estimated effect on traumatic brain injury is the presence of a skull fracture. Riders with a skull fracture had odds of traumatic brain injury that were more than 10 times greater than those of rid-

ANNALS OF EMERGENCY MEDICINE

41:1

JANUARY 2003

TRAUMA IN MOTORCYCLE RIDERS Kraus et al

ers with no skull fracture. This is true in both helmeted and unhelmeted riders. The adjusted OR for the presence of each facial bone fracture is smaller than the crude OR for the same bone. The mandible is significantly associated with traumatic brain injury only in the helmeted rider group (OR 3.76; 95% CI 1.98 to 7.15). The presence of a zygomatic fracture is associated with traumatic brain injury in both helmeted and unhelmeted riders. Product term ORs must be considered when using the table. Each facial fracture OR is relevant only in the absence of a skull fracture (and vice versa) because the facial fracture ORs and skull fracture ORs are not multiplicative. For example, in unhelmeted riders, the OR for mandible fracture in the presence of a skull fracture is 0.88 (1.10 × 0.80). A sparsity of data precluded the inclusion of a zygoma-skull fracture product term in the models. Table 5 shows results from logistic regression model 2. The probability of traumatic brain injury was modeled with upper facial bone fracture (orbit, nose, and zygoma), lower facial bone fracture (mandible and maxilla), skull fracture, and all product terms sepa-

rately for helmeted and unhelmeted rider populations. The estimated relation between facial bone fracture and traumatic brain injury is stronger for upper facial bones, regardless of helmet use status. In unhelmeted riders, the OR is higher for upper facial fractures (OR 3.40; 95% CI 1.99 to 5.82) than it is for lower facial fractures (OR 2.03; 95% CI 1.25 to 3.29). ORs for fractures of both the upper and lower facial bones were higher in the helmeted riders. The OR for upper facial bones is 1.75 times higher in helmeted riders (OR 5.94; 95% CI 2.68 to 13.14) than in unhelmeted riders (OR 3.40; 95% CI 1.99 to 5.82). ORs and CIs as predicted by model 2 are provided for riders with each of all possible combinations of facial fracture, skull fracture, and helmet use (Table 6). The table can be used to predict the OR for a theorized motorcycle rider with a given set of characteristics. For example, a rider presenting at an ED with indication of helmet use, a skull fracture, and an upper facial fracture is predicted to have odds of traumatic brain injury that are 12.0 times greater than those of a helmeted rider with no facial or skull fracture. Conversely, a helmeted

Table 3.

Number, ORs, and 95% CIs of traumatic brain injury in motorcycle riders with fractured facial bones by number of bones and fracture location, California, 1991 to 1993. Helmet Worn No. of Fractured Fracture No. With No. With Facial Bones Location TBI No TBI 1

≥2¶

Mandible Maxilla Nose Zygoma Orbit Mandible Maxilla Nose Zygoma Orbit

21 4 8 5 6 13 21 9 14 13

15 7 7 0 4 6 12 7 8 5

All Riders*

No Helmet Worn

OR†

95% CI

4.58 1.87 3.74 —ll 4.91 7.09 5.73 4.21 5.73 8.51

2.25–9.40 0.46–7.11 1.23–11.48 — 1.23–20.71 2.51–20.97 2.67–12.42 1.43–12.56 2.24–14.96 2.82–27.40

No. With No. With TBI No TBI 21 23 20 14 14 18 44 20 35 37

25 8 11 3 6 12 23 15 12 17

OR‡

95% CI

1.38 4.72 2.99 —ll 3.83 2.46 3.14 2.19 4.79 3.58

0.74–2.58 2.00–11.54 1.35–6.69 — 1.37–11.22 1.12–5.47 1.83–5.42 1.06–4.53 2.38–9.83 1.93–6.67

No. With No. With TBI No TBI 43 28 31 20 20 34 72 38 57 58

47 16 19 3 10 22 38 23 22 23

OR§

95% CI

2.71 5.19 4.84 —ll 5.93 4.59 5.62 4.90 7.69 7.48

1.75–4.21 2.70–10.07 2.64–8.94 2.64–13.61 2.59–8.14 3.71–8.53 2.83–8.53 4.57–13.01 4.49–12.54

*Includes

riders whose helmet use status is unknown. for OR is 642 helmeted riders with no facial fracture and TBI and 2,101 helmeted riders with no facial fracture and no TBI. ‡Reference for OR is 667 unhelmeted riders with no facial fracture and TBI and 1,096 unhelmeted riders with no facial fracture and no TBI. §Reference for OR is 1,356 riders with no facial fracture and TBI and 4,023 riders with no facial fracture and no TBI. llOR and 95% CI were not computed because of sparse data. ¶Fractures in 174 riders with ≥2 fractured facial bones. †Reference

JANUARY 2003

41:1

ANNALS OF EMERGENCY MEDICINE

2 3

TRAUMA IN MOTORCYCLE RIDERS Kraus et al

rider with a lower facial bone fracture only is predicted to have odds of traumatic brain injury that are 3.3 times greater than those of a helmeted rider with no facial or skull fracture.

fractured bones. Keenan et al 14 reported significantly increased risks (ORs) for traumatic brain injury among bicyclists with fractures of the orbit and maxilla but not for riders with fractures of the nose. The OR for traumatic brain injury was increased for frac-

DISCUSSION

The findings from this study expand on earlier reports that focused mainly on clinical descriptions of facial fractures and associated injuries. Despite the extensive clinical literature on facial bone fractures, few studies 11-14 have attempted to evaluate the risk of brain injury with concurrent facial bone fractures, and only one report has provided estimates of relative risk (ORs) of traumatic brain injury with specific

Table 5.

Crude and multivariate ORs and CIs for traumatic brain injury in injured motorcycle riders by presence of facial fracture and helmet use status, California, 1991 to 1993: Model 2. Helmet Use Status No

Facial fracture Upper Lower Skull fracture Product terms Upper × lower Upper × skull fracture Lower × skull fracture Facial fracture Upper Lower Skull fracture Product terms Upper × lower Upper × skull fracture Lower × skull fracture

Table 4.

Crude and multivariate ORs and CIs for traumatic brain injury in injured motorcycle riders by presence of facial fracture and helmet use status, California, 1991 to 1993: Model 1. Helmet Use Status No

Yes

Variable Facial fracture Mandible Maxilla Nose Orbit Zygoma Skull fracture Product terms Mandible × skull fracture Maxilla × skull fracture Nose × skull fracture Orbit × skull fracture Facial fracture Mandible Maxilla Nose Orbit Zygoma Skull fracture Product terms Mandible × skull fracture Maxilla × skull fracture Nose × skull fracture Orbit × skull fracture

*Adjusted

Crude OR (95% CI)

Adjusted OR* (95% CI)

1.54 (0.97–2.44) 3.29 (2.13–5.09) 2.27 (1.37–3.75) 3.33 (2.02–5.49) 4.93 (2.74–8.85) 9.10 (6.48–12.78)

1.10 (0.63–1.92) 2.08 (1.09–3.92) 1.80 (0.95–3.40) 1.14 (0.49–2.44) 2.63 (1.12–6.21) 10.68 (7.15–15.93)

— — — —

0.80 (0.19–3.39) 0.26 (0.08–0.80) 0.27 (0.07–1.00) 0.99 (0.27–3.64)

5.07 (2.92–8.80) 4.07 (2.23–7.44) 3.72 (1.83–7.59) 6.51 (2.93–14.45) 7.32 (3.19–16.80) 12.21 (7.56–19.74)

3.76 (1.98–7.15) 0.77 (0.28–2.09) 1.94 (0.82–4.56) 4.10 (1.33–12.64) 3.87 (1.25–11.99) 12.52 (7.31–21.45)

— — — —

0.45 (0.06–3.41) 0.95 (0.12–7.37) 0.63 (0.03–12.1) 0.09 (0.01–0.66)

Yes

Crude OR (95% CI)

Adjusted OR* (95% CI)

3.74 (2.59–5.39) 2.53 (1.80–3.55) 9.10 (6.48–12.78)

3.40 (1.99–5.82) 2.03 (1.25–3.29) 10.38 (6.91–15.59)

— — —

0.43 (0.17–1.06) 0.59 (0.20–1.74) 0.37 (0.14–1.04)

5.54 (3.33–9.22) 4.42 (2.86–6.85) 12.21 (7.56–19.74)

5.94 (2.68–13.14) 3.29 (1.81–5.97) 12.03 (7.03–20.60)

— — —

0.23 (0.09–0.78) 0.02 (0.03–0.91) 1.25 (0.14–8.04)

*Adjusted

for facial fracture location, presence of skull fracture, and facial fracture–skull fracture interactions.

Table 6.

for facial fracture location, presence of skull fracture, and facial fracture–skull fracture interactions.

2 4

Variable

Predicted ORs for riders with selected characteristic combinations from logistic regression analysis of injured motorcycle riders, California, 1991 to 1993. Helmet Use Status Unhelmeted Helmeted

*Reference †Reference

Facial Fracture Location

No Skull Fracture, OR (95% CI)

Skull Fracture, OR (95% CI)

None Lower Upper None Lower Upper

1.00* (—) 2.03 (1.25–3.29) 3.40 (1.99–5.82) 1.00† (—) 3.29 (1.81–5.97) 5.93 (2.68–13.14)

10.38 (6.91–15.59) 7.88 (3.10–20.02) 20.74 (7.70–55.84) 12.03 (7.03–20.59) 49.34 (7.59–320.59) 11.98 (2.53–56.85)

group is unhelmeted riders with no facial fracture and no skull fracture. group is helmeted riders with no facial fracture and no skull fracture.

ANNALS OF EMERGENCY MEDICINE

41:1

JANUARY 2003

TRAUMA IN MOTORCYCLE RIDERS Kraus et al

ture of the mandible, but it did not quite reach statistical significance, likely because of very small sample sizes. Our findings show significantly increased odds of traumatic brain injury with fractures of any facial bone, with odds varying by the bone fractured. Most commonly, only a single facial bone is fractured in motorcycle crashes; the highest chances of traumatic brain injury are associated with fracture of the zygoma or orbit, and the lowest odds (still significantly increased) are associated with fracture of the mandible. Odds of traumatic brain injury varies also by whether 1 or more facial bones are fractured. Although odds of traumatic brain injury could not be derived for riders with more than 1 fracture for each bone, it is noteworthy that approximately 15% of riders with a fracture of the mandible or maxilla and approximately 10% of riders with fracture of the orbit have multiple fractures to each of these bones. Our findings indicate that facial fractures are a good indication that brain-damaging energy transfer might have occurred, and at the very least, an impact to the head occurred. Helmets, which reduce this energy transfer, were found to be protective of traumatic brain injury in our multivariate analyses. However, in helmeted riders the facial fracture–traumatic brain injury association was somewhat higher than that seen in unhelmeted riders. This finding might indicate that when a facial fracture occurs and a helmet was worn, the effect might have been severe enough and the energy transfer great enough to cause traumatic brain injury. Therefore, clinicians should evaluate traumatic brain injury when patients wearing helmets present with facial fractures, especially those of the upper face. Although motorcycle riders with fractures of the lower facial bones had increased odds of traumatic brain injury, this outcome was most strongly linked to fractures of upper facial bones with or without fractures in the lower face. Facial injuries are a common occurrence in motorcycle crashes, and fractures are diagnosed in a substantial proportion of those with visible damage to the face. This fact alone, irrespective of concurrent brain injury, warrants efforts to protect the face and head of riders exposed to the risk of a motorcycle crash. That soft tis-

JANUARY 2003

41:1

ANNALS OF EMERGENCY MEDICINE

sue injury to the face is related to an increased risk of traumatic brain injury plus the psychologic effect of severe facial injury21,24 suggests that preventive measures are needed to mitigate this form of injury. Most reports11-14 agree that the lowest risk of traumatic brain injury is associated with fracture of the mandible (lower face), and the highest risks are associated with fracture of all bones in the upper face, except the nose. This finding has implications for helmet type and mandatory helmet use for motorcyclists. There is no question that helmets reduce the risk of traumatic brain injury in crashes, and the earlier work of several researchers3,6 underscores the importance of full-face helmets in operating motorcycles or scooters. There is wide variation in human tolerance to facial bone fracture. Hampson,25 in an extensive review of published data on facial bone fracture pressures, shows wide ranges in energy levels necessary to fracture specific bones of the face. The data from cadaver samples in the Hampson report might not be the same for living persons. Nonetheless, there does not appear to be a correlation between levels of fracture pressures and risk of traumatic brain injury. For example, the nasal bone with lowest fracture pressures (and less implied energy transmission) has an intermediate risk of traumatic brain injury, and the zygoma with an intermediate fracture pressure level has the highest risk estimate of traumatic brain injury. Model 2 (Table 6) is useful in understanding the interactions among facial fractures, skull fractures, and helmet use status. The effect of the interaction between upper and lower facial bone fractures can be observed in the OR for riders with fractures of both the upper and lower face. For example, in unhelmeted riders without a skull fracture, the OR would be 6.8 in the absence of interaction. With the observed interaction, the predicted OR is 2.9. In riders with a skull fracture, the interaction is more difficult to visualize. The relatively small OR of 6.7 that the model predicts for unhelmeted riders with fractures of both the upper face, lower face, and the skull is a reflection of the large coefficients for the upper × lower, upper × skull, and lower × skull product terms.

2 5

TRAUMA IN MOTORCYCLE RIDERS Kraus et al

There were several limitations in this study concerning using and coding existing hospital record data. Perhaps not all medical diagnoses were recorded, and some might have been miscoded. The Abbreviated Injury Scale does not allow detail in diagnostic coding; for example, supraorbital ridge is subsumed under orbit. These shortcomings and occasional small sample sizes prohibited more detailed analyses. Some investigators11,21 have suggested that facial bone fractures lessen the risk of traumatic brain injury because they act as an energy-absorbing cushion and thus protect the intracranial organs. Findings of this study and those of other investigators13,14 do not support this assertion. In fact, we provide evidence of an opposite effect, namely facial bones when impacted transmit energy forces to the brain. This is seen in the high odds of traumatic brain injury in riders without a skull fracture but with fractured bones of the face. The findings of this study support previous research demonstrating an association between facial fractures and traumatic brain injury. ED practitioners should screen all motorcyclists with facial fractures for brain injury, regardless of helmet use status. Increased use of full-face motorcycle helmets might decrease the risk of traumatic brain injury occurrence by reducing energy transfers to the facial bones of crash-involved motorcycle riders.

REFERENCES

Author contributions: JFK conceived, designed, and executed the original study. CPA had day-to-day project management responsibility, including data management. DLM performed some of the preliminary data analyses. JFK drafted the manuscript, and TMR and CPA performed the final analyses and created the final drafts of the manuscript. JFK takes responsibility for the paper as a whole.

19. Down K, Boot D, Gorman D. Maxillofacial and associated injuries in severely traumatized patients: implications of a regional survey. Int J Oral Maxillofac Surg. 1995;24:409-412.

Received for publication March 8, 2001. Revisions received August 13, 2001, and November 30, 2001. Accepted for publication December 17, 2001. Supported by grant No. R49/CCR903622 to The Southern California Injury Prevention Research Center from the Centers for Disease Control and Prevention and the UCLA Brain Injury Research Center. Address for reprints: Jess F. Kraus, MPH, PhD, The Southern California Injury Prevention Research Center, UCLA School of Public Health, 10911 Weyburn Avenue, Suite 200, Los Angeles, CA 90024; 310-794-2706, fax 310-794-0787; E-mail [email protected].

1. Nakhgevany K, LiBassi M, Esposito B. Facial trauma in motor vehicle accidents: etiological factors. Am J Emerg Med. 1994;12:160-163. 2. Sastry S, Sastry C, Paul B, et al. Leading causes of facial trauma in the major trauma outcome study. Plast Reconstr Surg. 1995;95:196-197. 3. Cannell H, King J, Winch R. Head and facial injuries after low-speed motorcycle accidents. Br J Oral Surg. 1982;20:183-191. 4. Gopalakrishna G, Peek-Asa C, Kraus J. Epidemiologic features of facial injuries among motorcyclists. Ann Emerg Med. 1998;32:425-430. 5. Koorey A, Marshall S, Treasure E, et al. Incidence of facial fractures resulting in hospitalisation in New Zealand from 1979 to 1988. Int J Oral Maxillofac Surg. 1992;21:7779. 6. Bjornstig U, Ostrom M, Eriksson A, et al. Head and face injuries in bicyclists with special reference to possible effects of helmet use. J Trauma. 1992;33:887-893. 7.

Gassner R, Hackl W, Tuli T, et al. Facial injuries in skiing. Sports Med. 1999;2:127-134.

8. Gruss J, Pollock R, Phillips J, et al. Combined injuries of the cranium and face. Br J Plast Surg. 1989;42:385-398. 9. Cannell H, Silvester K, O’Regan M. Early management of multiply injured patients with maxillofacial injuries transferred to hospital by helicopter. Br J Oral Maxillofac Surg. 1993;46:207-212. 10. Assael L. Clinical aspects of imaging in maxillofacial trauma. Radiol Clin North Am. 1993;31:209-220. 11. Lee K, Wagner L, Lee Y, et al. The impact-absorbing effects of facial fractures on closed-head injuries. J Neurosurg. 1987;66:542-547. 12. Davidoff G, Jakubowski M, Thomas D, et al. The spectrum of closed-head injuries in facial trauma victims: incidence and impact. Ann Emerg Med. 1988;17:6-9. 13. Haug R, Adams J, Conforti P, et al. Cranial fractures associated with facial fractures. J Oral Maxillofac Surg. 1994;52:729-733. 14. Keenan H, Brundage S, Thompson D, et al. Does the face protect the brain? Arch Surg. 1999;134:14-17. 15. Vaughan R. Motor cycle helmets and facial injuries. Med J Aust. 1977;1:125-127. 16. Johnson R, McCarthy M, Miller S, et al. Craniofacial trauma in injured motorcyclists: the impact of helmet usage. J Trauma. 1995;38:876-878. 17. Lee M, Chu W, Chang L, et al. Craniofacial injuries in unhelmeted riders of motorbikes. Injury. 1995;26:467-470. 18. Haug R, Savage J, Likavec M, et al. A review of 100 closed head injuries associated with facial fractures. J Oral Maxillofac Surg. 1992;50:218-222.

20. Haug R, Prather J, Indresano A. An epidemiologic survey of facial fractures and concomitant injuries. J Oral Maxillofac Surg. 1990;48:926-932. 21. Lim L, Lam L, Moore M, et al. Associated injuries in facial fractures: review of 839 patients. Br J Plast Surg. 1993;46:635-638. 22. Association for the Advancement of Automotive Medicine. The Abbreviated Injury Scale. Des Plaines, IL: AAAM; 1990. 23. Hosmer DW, Lemeshow S. Goodness of fit tests for the multiple logistic regression model. Commun Stat. 1980;A10:1043-1069. 24. Bisson J, Shepherd J, Dhutia M. Psychological sequelae of facial trauma. J Trauma. 1997;43:496-500. 25. Hampson D. Facial injury: a review of biomechanical studies and test procedures for facial injury assessment. J Biomech. 1995;28:1-7.

We thank A. A. Afifi, PhD, for statistical consultation.

2 6

ANNALS OF EMERGENCY MEDICINE

41:1

JANUARY 2003